I've uploaded the NGC 7790 standards field photometry from my calibration work at USNO-Flagstaff using the 1.0m telescope. There are something over 300 stars now available, ranging from 11<V<18, and with all UBVRI filters.
From 11<V<14, the data comes from short exposures of the field, designed to not saturate the three cepheids in the cluster. There are about 11 nights of data that were of proper quality. For 14<V<18, the photometry is from longer exposure images on 7 nights.
There are a few complications. First, there were a handful of existing comparison stars in the field. While their photometry was updated with this import, and they are of standards quality, they will not show up when you click "Standards" in the VSP input form. Right now, a star is either a comparison star, or a standard. We will have to resolve that in the next VSP release. However, if the star is listed with a sourceID of 10, it is of the same quality as the others in the field and can be used for determining your coefficients.
Second, there are two NSV stars in the cluster. NSV 11781 is a delta Scuti variable with small amplitude, but I've removed it from the standards file. Note that the current VSX coordinates are off by a couple of arcsec. NSV 26181 is perhaps a very low-amplitude red variable (also with poor VSX coordinates). I've left it in the standards file because it is also the reddest star in the field.
To access this standards field, enter the coordinates 23:58:23.2 +61:12:25 in the VSP form, and click "would you like a standard field chart?".
Enjoy!
Three cheers for our Director: Hooray, hooray, hooray: Wow, what a lot of work! Thank you, thank you, thank you, Arne!
Hello Arne
This is great news. Any chance of breaking the rules a little, and making it so that VSP recognizes "NGC 7790" as a "program star" and gives you the info. That way one does not have to google the coordinates of the cluster to get the info?
It would be great to do the same for M67, and the other Standard Clusters that folks are going to use a lot. There are probably a dozen fields that AAVSO Observers are going to be expected to use a lots. Perhaps six of these are Landoldt Fields. I realize that HQ would have to draw the line somewhere, and folks could think of hundreds of fields, but that we don't want. How about a Dozen--at your choise?
Gary
Gary,
If you type CE Cas you can print the chart of the NGC 7790 standard field. It works for me.
Clear skies,
Robert
Hooray and thank you, Arne and Sara!
Another way to bring up NGC 7790 is to use these chart I.D.'s in the VSP "Advanced Options", "Do you have a Chart ID?" box.
Chart 13611ASR for a 15 min chart, Chart 13611ASS for a 10 min chart, Chart 13611ASV for a 7.5 min chart
Is anyone else having problems printing the photometry table? On eleven printed pages I only get the leftmost 6 columns.
Phil
Hello Phil
I too was having problems printing. I just went down to the local library and had no problems. Thought it was something on my end. Maybe not.
The point of being able to get the chart and the photometry table, by inputting "NGC7790", is that's something I can remember without additional steps. I can always pull up Simbad, and get the Ra and Dec, write it down, or copy and paste. Just thought in this day and age, AAVSO should allow for about a dozen standard clusters by name, the ones that folks would be using regularly.
BTW: Of more interest is that the values in the NGC 7790 "New" sequence, differ from one I pulled up about a month ago on VSP. Some stars are different by 0.2 to 0.3 mags. I plan to recalculate my Transformation Coef and see what the impact is. I have been using the Transformation coef that were calculated using the VSP values from a month ago. I also noticed that the SWRI values published are different from the VSP values of a month ago.
Gary
About the suspected variables, the brighter one is NSV 14781 not 11781 and it is an early B-type (B1-B2) star that was found showing emission at least once in the past so it might be a Be star.
Its coordinates have been corrected in VSX.
I noticed that the red one is found to have V= 12.89 from USNO-Flagstaff which is fainter than the literature range (12.66 - 12.82) so maybe it is in a fainter state now. It will be interesting to see what the observers find.
Cheers,
Sebastian
-----------------------
Sebastian Otero
VSX Team
American Association of Variable Star Observers
Are these revised standard values the result of recent measurements that Arne made with the USNO 1.0M telescope or are they just new to AAVSO. Cause of my question was Sebastian's statement that NSV 26181 might be in a fainter state now. That would only be applicable if these measurement were made recently as well as being new to AAVSO.
Brad Walter
Hi Gary,
Unfortunately, I wiped out the existing magnitudes without writing them down, so I was wondering if you had a table printout with the old values that you could scan, so I can see where the differences occur. Those may have been added from Seqplot fairly recently, and unfortunately for this field, Seqplot doesn't work well because of the short- and long-exposure sets.
I will look into plotting a few fields by name, even if it isn't a variable star. As mentioned earlier, you can use one of the known variables in the cluster as a quick-and-dirty way to get a chart, such as CF Cas.
Arne
Arne, and Gary
Is the attached the OLD NGC 7790 SEQUENCE you are looking for? It is the one I have been using for the last few years. I have attached it as a PDF with the field image and photometry sequence and the sequence itself as an Excel file.
To: Arne,
I will verify that the camera is producing files with accurate filter designations. I'm pretty sure that they are since I've made no configuration changes since I took these data on EE Cep:
I will get more stars to run in AIP4WIN to see if the T.C.'s change.
James
Well, I haven't started using the new tools yet, so I cranked through spreadsheet calculations in a plodding old fashioned way, but one that let's me see what I am doing step by step through the process. I also graph each of the transformation data sets and check to see if I have any outliers. that should be eliminated. I also used all 31 stars with sequences form the 33 stars labeled on the OLD NGC7790 standard field chart that I posted to this forum thread earlier.
The attached spread sheet has two sets of IRVBBVRI image sequences. The first set was centered at airmass 1.67 and the second at airmass 2.43. I did calculations on the first set by itself and the two sets together (as a double set of measurements) just to see what affect changing airmass would have. It had a significant effect at these airmasses. I used the data from set 1 (lower airmass) in my transformation calculations. You can ignore the data fand calculations or the second set.
All measurements are raw instrumental magnitudes. That is easy to do with the software I use. I could have used a comp star since it should only affect the offset, not the slope. I calculated the slopes of V-v to B-V and similar slopes, C-c vs C for the color indexes B-V, V-R, V-I and R-I. The slope of V-v vs. B-V regression line is Tv but for the color indexes, if Mc is the slope of the X-Y plot regression line then Tc = 1/(1-Mc). I could have taken the slopes right off the plot legends but I calculated the LINEST regressions on the NGC 7790 Data tab because I wanted standard errors for the slopes to calculate standard errors of the transformation coeficients. The results from the this night' and two other nights and the averages of all three is presented on the Transformations tab.
This looks like a bunch of work but once you have the spreadsheets set up it isn't that bad and going through it step by step a few times allowed me to really understand what was going on.
I found that in all my set 1 measurement data on all nights, except for Tri on one night (this one) where some "stuff" came through briefly, The Tv data had the highest residuals around the regression line followed by Mbv. I would expect Tv to have the worst fit since the slope should be nearly zero (little or no correlation). The residuals for Mvr, Mvi always had less scatter. In some cases the fit to the regression line was extremely close. The Tv data never really looked as though it approximated a line and Mbv and Mri required varying amounts of imagingation from night to night.
When making the photometry measurements, I found I had to be very careful to check the positions of the measurement apertures. Several of the stars have close companions that can cause the aperture to shift a bit relative to the real centroid of the star if you aren't careful. It is tedious, but for all stars in all images, I checked the postion of every measurement aperture to make sure all were located on the centroids. I think I now have the postions and numbering of these stars memorized, at least for the old sequence numbering scheme.
So although this claculation is a bit long-hand (Do I I hear Gary groaning in the distance) I hope seeing a set of calculation worked through and some transformation values from a typical amateur imaging set up may be useful..
Brad Walter
Hello Brad
This sequence looks fine. The mistake that I made, was that I querried VSP for a chart of NGC7790, and got the standard AAVSO chart and sequence. Some of the sequence values did not agree with the "binaries-boulder-swri.....) sequence. This was the difference. That has since been changed, so now VSP and "binaries-boulder-swri.....)" do agree.
And yes Brad, I too have a spreadsheet to determine Transformations that I did again before the spring meeting when I found out about Gordon's effort. And yes, that was a groan, you heard. I too like looking for outliers.
As a final test of the end to end process, I observed SA 110-503 and averaged the result over 2 night, about 30 data points in each color. I compared to Landolt. I was -3, 0, +6 and 0 millimags (BIRV) from the published values. The errors on the published values were 4, 4, 4, & 3. My Std Dev's on the best night were 6, 8, 7, and 4.
BTW: I am using TA to apply the coeficients and its working very well.
Gary
That and I was also recalling the story you told me about a prof. making you calculate and verify the transformation coefficients for every combination of filters in the J-C "coloring book" even though you would probably never use several of them.
I realize that I used the "traditional" photomtric reduction method detailed in Henden and Kaitchuck. Brian D Warner's method in A Practical Guide to Lightcurve Photomoetry and Analysis would have worked just as well. This is just the way I am used to doing it. I know I will end up using the new AAVSO tools soon. My aging pooch has taught me that the saying "you can't teach an old dog new tricks" is wrong. He quickly learns the new ones when they require less effort and yield the same or greater quantity of treats.
Brad Walter, WBY
Thanks Arne,
•Prepared VPHOT seq. Limited magnitudes equal to or less than 13.5 in V
•28 Stars, B-V range: 0.378 to 1.742
•Did notice a couple of arcseconds of coordinate errors between Plate Solved VPHOT image and the table data as noted
•Star# 131 000-BLJ-976 produces significant error 0.191 vs 0.041 in the Target measurements, is this a table typo error?
•Would like to share VPHOT sequence to obtain a check on typo/ID-coord selection errors
Mike
I've picked 3 stars from the standards list to compare with the compilation created by Peter Stetson:
http://www4.cadc-ccda.hia-iha.nrc-cnrc.gc.ca/community/STETSON/standard…
ID RA DEC V B-V SID SV SBV
000-BLJ-965 359.50851 61.19128 12.488 0.023 0.537 0.035 N7790-S26 12.471 0.006
000-BLJ-987 359.74005 61.16264 13.336 0.029 1.489 0.045 N7790-S402 13.344 0.007 1.485
000-BLK-044 359.50140 61.23952 14.906 0.021 0.435 0.032 N7790-S16 14.902 0.006
where SID = Stetson's ID, SV is V in his table, SBV is (B-V) in his table (where present).
As you can see, the current VSP measures match literature values to within statistical error. The good features is that they are all taken with the same system and so are homogeneous.
If the half-dozen comparison stars in this field were 0.1-0.2mag offset from the current calibration, then the old values were wrong (and again, probably drawn from Seqplot where this particular field is pathological because of the short/long merged exposure datasets).
Arne
Mike asked
Star# 131 000-BLJ-976 produces significant error 0.191 vs 0.041 in the Target measurements, is this a table typo error?
Looking at Stetson's table, this is star N7790-S62, which has V = 13.109 +/- 0.004. My photometry yields 13.104, so the two measures agree; the current VSP is correct. This is a pretty red star, so you may have to transform before comparing numbers closely.
Note: Stetson reports standard error of the mean. I report standard error of a single measure, and so overestimate the amount of error in my photometry. Consider my uncertainties as being very conservative.
Arne
I think this is a great initiative. It would be good though to call all of the Landolt fields that are listed here http://www.aavso.org/standard-stars-vsp standard stars. They are all great to help measure transform values.
Terry
Good point Terry,
The Cousins E and F Region standard stars should also be included. Cheers,
Mark
Do these numbers look good?
I shot 16 images of NGC 7790 in BVRI on 30-Aug-14. I got the BVRI data using 9 comparison stars exported from the VSP photometry chart of CE Cas using PhotomCap_14. This translated the photometry table into a readable data file that AIP4WIN uses. I then brought-up all my image files in AIP4WIn and processed them using magnitude measurement tool; I used a zero point of 20.36 to get non-negative magnitude values. After tagging all 9 comparison stars in the NGC 7790 field, I ran the instrumental magnitude report from AIP4WIN 2.4.8. I averaged the values of the first 3 of each 4 BVRI set using the comparison stars instumental magnitudes; I used the middle (of each set) for the Air mass value.
It took me a little while to make PTGP to work on my data, because I used the AUID names from VSP. I had to translate these into the NGC 7790 Standards Field ID number that is used in the PTGP to determine the TC's. If this looks good, I'll be ready to join the "fun" of taking and comparing TC data for XZ Cet.
James
Hello James
What I did, actually am still doing to test out trnasformations, TA and my Trans Determiner, and PT software, is that I image a known Landolt Cluster and see what I get for Transformed Values. The TA team has used SA 110-503 as a target. Right now I am within 1, -1, 10 and 13 mmags for BVRcIc. With prior coeficients, I was 20-86 mmags off. Another member of the TA team is within 10 mmags also. I used 3 nights, 20-35 images per night per color.
You can use any good cluster, it does not have to be SA 110. Choose one at high air mass at the time of your observatons.
I have attached a Summary Table and a plot of the data. The symbols are purposely small so that the error bars show up.
Gary
Hello James
Could you post the 9 stars you used and the values that you used? I have Tx coef with a good set and a better set. I many be able to see if your comp star values are the new ones or old ones. If you go to VSP, you should be getting the new ones now. There was apparently some updating between 2 July and 31 Aug.
I plan to redo my Tx coef with >30 or so stars, just to see what a difference that makes from the better set. FYI, my B-V was from 0.2 to 0.7 mag. It looks like I can extend that to 1.5 to 2.0 on the red end. I will let you know what I find.
Gary
Hi James,
These coefficients don't look right. Tv(BV) for example is 0.358; for most CCD systems using a commercial V filter, that coefficient is typically in the -0.1 < Tv(BV) < 0.1 range. The only magnitude coefficient that will be outside of that +/- 0.1 range might be Tb, depending on your filter and your CCD as many CCDs don't have good blue response. My guess is that some part of this difference is due to the restricted color range of the standard stars that you are using.
When displaying the field using a variable-star name, you need to be sure that you go to the bottom of the VSP form and click "Yes" for "Would you like a standard field chart?". That will bring up all 300+ standards for the field. If you want, you can restrict the stars that you measure to a smaller subset, but be sure to include most, if not all, of the very blue and very red stars.
Gordon is working on revising PTGP to use AUID's, so some of the extra steps that you had to take will go away.
Arne
To: Arne/Gary,
Thanks for the feedback! Here are the 9 comparison stars used to get my transforms (from The CE Cas chart):
AUID
RA.
Dec.
Label
B
V
B-V
Rc
Ic
000-BFX-116
23:58:38.21 [359.65921d]
61:19:19.5 [61.32208d]
110
11.474 (0.030)10
11.017 (0.026)10
0.457 (0.040)
10.754 (0.027)10
10.453 (0.027)10
000-BFX-118
23:57:49.15 [359.45480d]
61:19:44.3 [61.32897d]
117
11.875 (0.079)10
11.659 (0.078)10
0.216 (0.111)
11.555 (0.080)10
11.392 (0.080)10
000-BFX-119
23:57:27.77 [359.36572d]
61:06:24.5 [61.10680d]
119
12.363 (0.019)10
11.944 (0.017)10
0.419 (0.025)
11.667 (0.023)10
11.323 (0.023)10
000-BFX-120
23:57:35.15 [359.39645d]
61:19:08.8 [61.31911d]
122
12.801 (0.043)10
12.244 (0.041)10
0.557 (0.059)
11.904 (0.042)10
11.557 (0.042)10
000-BFX-123
23:58:03.11 [359.51297d]
61:08:14.1 [61.13725d]
124
13.141 (0.031)10
12.452 (0.027)10
0.689 (0.041)
12.031 (0.029)10
11.598 (0.035)10
000-BFX-124
23:58:13.76 [359.55734d]
61:10:36 [61.17667d]
126
13.063 (0.031)10
12.617 (0.029)10
0.446 (0.042)
12.363 (0.032)10
12.056 (0.041)10
000-BFX-125
23:59:14.06 [359.80859d]
61:09:17.1 [61.15475d]
128
13.612 (0.020)10
12.806 (0.017)10
0.806 (0.026)
12.323 (0.023)10
11.811 (0.034)10
000-BFX-127
23:58:48.76 [359.70316d]
61:12:34.6 [61.20961d]
131
13.667 (0.021)10
13.172 (0.018)10
0.495 (0.028)
12.892 (0.022)10
12.549 (0.032)10
000-BFX-128
23:58:56.22 [359.73425d]
61:14:54.6 [61.24850d]
145
15.077 (0.147)10
14.467 (0.127)10
0.610 (0.194)
14.078 (0.131)10
13.626 (0.131)10
I notice that the B-V for these are towards the red with the bluest star 000-BFX-118 at B-V=0.216.
Regarding: "Tv(BV) for example is 0.358; for most CCD systems using a commercial V filter, that coefficient is typically in the -0.1 < Tv(BV) < 0.1 range. "
I thought this was high as well. My B & V filters are astrodon photometric dichromatics, but since my selection of stars were limited and my average B-V value for these is 0.522, these stars might have caused this. One thing to note, I'm taking data from extremely light polluted skies. My average backround ADU for these 180sec 1x1 bin exposures of NGC 7790 are as low as 500 for the I filter and as high at 2200 for the R & V filter; the B is at about 1500.
I've done a couple of runs from a dark sky site, 90 miles NW of LA from a 2500 meter mountain and my backround ADUs were 1/10 these values.
James
Hello James
I am happy to report that all 9 of these stars match the PT chart 13611AHV (revised values), which I pulled off the web on 31 August and printed out. That is not the cause of your difficulties.
I do suggest that you use a wider color range as Arne suggested. You can get from 0.216 to 2.0 in this field. Perhaps this will change your Tx coef. Use the "Standard Star Option". Perhaps increase the number to 30 or so?
Because I have been able to get good transformations with 6 of these stars, I suspect that the color range is not your only problem. You might carefully look at your pipeline, it might have something amiss.
I am in the process of culling 30 stars from this field, and recomputing my Tx coef. Let you know how this goes.
Gary
To: Gary,
Thanks for the confirm! It very well might be possible that something in my pipeline can be the issue. It was very hard to get the format correct in the way that PTGP4.0 would accept and transform. Here is the pipeline:
MaximDL5.23
Image,flat,dark,bias acquisition & calibration---->
AASVO VSP CE Cas Photometry Table---->
PhotomCap_14(Process CE Cas Photometry Table into .STAR file for AIP4WIN use---->
AIP4WIN 1.Open 12 images (3x BVRI)---->
AIP4WIN 2. Use MMT tool (zero pt=20.36) & tag the CE Cas variable V & 9 comp stars as c1, c2, etc---->
AIP4WIN 3. Run the Raw Instrument magnitude report. On exit of AIP4WIN save as Log.txt file---->
Open Log.txt and make it into format that is used by PTGP4.0. While doing this rename AUID names to ID names for the NGC 7790 Standard filed---->
Load this modified log.txt file into PTGP4.0 and get the Transform Coefficients(?)
Note that i used the demo M67 .txt file in PTGP4.0 as a format model for this log.txt data of NGC 7790. I used the BVRI magnitudes left for M67 on this file and just renamed it NGC 7790 its star data line; 1st object before the 9 compstar data. When I get home I can list this .txt file out. I manually formated this log.txt file because I could not get multifilter data into AIP4WIN without using PhotomCap_14 to format a .STAR data file for all the UVBRI magnitude of the 9 comp stars used.
James
To: Arne/Gary,
Here is the input file I used to get my transforms:
AIP4Win v2.4.8 Magnitude Measurement Tool
Instrumental Magnitudes Report
Format: One line per star. Images in columns.
Date: JD = Geocentric Julian Day.
Airmass: Computed for entire image.
Decimal places normal: Three for magnitude, five for date.
Delimiting Character: ;
Aperture = 5.00 Inner/Outer Annulus = 7.00/10.000
Observer name = FJQ at LON=-118.16944 LAT=34.03222
Target name = NGC 7790 at RA=359.5394 DEC=61.21367
Results computed on the basis of user input.
Now = 9/1/2014 10:00:31 PM
Image # ; 00001 ;; 00002 ;; 00003 ;; 00004 ;;
Julian_Day ;2456899.91447 ;;2456899.93406 ;;2456899.94383 ;;2456899.92439 ;;
Integr ; 180 ;; 180 ;; 180 ;; 180 ;;
Filt ; B ;; V ;; R ;; I ;;
Airmass ; 1.1316 ;; 1.1465 ;; 1.157 ;; 1.1382 ;;
C1 Sky ; 1266.63 ;; 2155.81 ;; 2588.06 ;; 641.22 ;;
ADUmax ; 26788 ;; 53449 ;; 57152 ;; 50115 ;;
NGC 7790; 12.652; 0.006; 11.955; 0.006; 11.335; 0.004; 11.493; 0.005;
110; 11.436; 0.001; 10.979; 0.001; 10.793; 0.001; 11.725; 0.001;
117; 11.863; 0.002; 11.587; 0.001; 11.532; 0.001; 12.597; 0.002;
123; 12.529; 0.002; 12.078; 0.002; 11.890; 0.002; 12.780; 0.002;
122; 12.714; 0.003; 12.167; 0.002; 11.907; 0.002; 12.794; 0.003;
127; 12.993; 0.003; 12.360; 0.002; 12.027; 0.002; 12.799; 0.003;
128; 13.065; 0.003; 12.635; 0.003; 12.465; 0.003; 13.394; 0.004;
130; 13.443; 0.004; 12.726; 0.003; 12.339; 0.002; 13.038; 0.003;
134; 13.632; 0.005; 13.137; 0.004; 12.950; 0.004; 13.832; 0.005;
145; 14.863; 0.013; 14.254; 0.010; 13.994; 0.009; 14.764; 0.009;
Please note that the target coordinates for NGC 7790 in the AIP4WIn header are for CE Cas. Also the magnitudes/error in the NGC 7790 line for B V R I are M67's magnitudes from the demo file in PTGP4.0, filename: M67_T17_20140330_Photometry_Demo_instrument_mags.txt.
I don't know how this was derrived for M67 let alone how to get NGC 7790's BVRI magnitudes. I assume PTGP uses this line for object field identification only.
James
James,
I ran your data through PTGP 4.0 and looked at the plots. Everything matches your results and no individual observations jump out as problems when looking at the plots. But I agree with Arne; I haven't seen coefficients as large as yours. One suspect is the comp stars. I just updated and will release as soon possible (after a few people test it) a PTGP update that will allow yiou to use AUID's (or Arne's original 1-64) identifiers. And I'm trying to figure out an approach to accept VPHOT.
Gordon
Hi James,
Remember that a coefficient of 0.358 means that for a star of pretty normal color, (B-V) = 1.0, you have to dim your measurement by 0.358 in order to match the standard system. That means your filter is much redder than the standard one. If we assume the transformation calculation is correct, then my guess is that this filter is actually an R or Sloan r' filter, not a V.
So I'd suggest several things. First, what filters are in your filter wheel? Only BVRI, or do you have a mix of other filters? Second, you are using a very restricted set of stars with small color range for your transformation. Do your calculation again, but include some of the red stars (there are several with (B-V) in the 1.5 range). If you get the same result, then I would next pull the camera, take out the filter wheel, and look through the filters. V ought to look like a pretty green. Make sure the order is what you expect. Plug the filter wheel into the computer, and have it select V, and make sure it really does this.
So start with a wider color range for your calculation, and let us know what you get.
Arne
To: Brad & Gary,
I appreciate the expertise you bring to the table with regards to your transform coeffiecents spreadsheets. I'm still waddling along but at the last Meeting in June, Dr Arne mentioned that AASVO is introducing tools to make calculating transform coeffiecents easier and more standard between observers. I'm anxious to start transforming my BVI data I take nightly to meet Arne's Pro standards. I hope to get within +/- 0.02 BVI magnitudes of other observers transformed magnitudes for XZ Ceti.
Back to NGC 7790. Per Arne's directions, I downloaded all UBVRI data to magnitude 16.5 from the RA/Dec coordinates 23:58:23.2 +61:12:25 in the VSP form, with the standard field chart option on. I also made a 300DPI resolution chart of this cluster to properly identify the stars to be taged in AIP4WIN to determine their instrumental magnitudes. My main difficulty with this is that more than 50% of the comp stars shown have multiple stars with the same label; one of them has 10 stars with the same magnitude label! In order to select the "right" star, I downloaded the list to excel and 1st selecting the stars with unique labels (highlighted yellow). I next selected stars with only 2 stars sharing the same label, then selected stars with only 3 stars sharing the same label. After selecting 3-4 sets of stars with no more than 4 similar labels, I had about 22 comp stars selected. From these binned (2-4 stars with the same label) I selected one from each set based on a low B-V value. Since this cluster has an extinction reddening of E(B – V ) = 0.51 blue stars are under represeted there.
I should be able to get a new set of transform coeffiecents after I've parsed these stars out and labeled them on my NGC 7790 chart. I always have discreate comp stars identified on the VSp chart before tagging them in software, especially if they have the same label and needed to be specifically identified so as not to apply the BVRI data to the wrong star; I wish the comp star labels had an extra digit to make this easier!
James
P.S. Arne, I took my CCD off the scope and made 600sec exposure of B,V, & R filter to view what filter was selected by MaximDL. I used a flashlight to verify that the camera correctly selected the physical filter selected triggered by MaximDL software when exposing, so this is not my issue with having a large Tv(BV) for Transform coefficient of 0.358. Will try to derrive new ones with a larger set of 21 (bluer) stars.
Hello James
I feel your pain. Was there last week. Too many stars, some with duplicates of 10 of the same label. Its a good thing for Professional Software scripts that cull things in cyberspace, but it does not work well for us Maxim and equivalent users. Some day, HQ may put a pipeline together that does all that with Arne's tools, but until then, we need Plan B.
What Arne suggested, and it worked pretty good for me, is go to the (binaries-boulder-7790-standards...google it and it will come up. There is a nice chart there, that works well for my fov, 31 stars are designated, there are no duplicates, and the PT in 4 colors and their errors are on the next page. There are B-V stars from 0.252 to 2.064 and lots in between. I use Maxim, and I have to point and click for each star. So I need some breathing room, and there are some choices where you can get one or the other of a pair. Choose the one that gives you the color you need. If you can get all 31 stars, great. These values do agree with VSP PT Chart if you access it now. So the numbers are good ones.
While not a HQ approved method, what I did was try 6 and then 13 stars from the SWRI list. I used SWRI# 8, 3, 30, 26, 23, 7, 13, 12, 1, 27, 24, 31 & 16. Its a pretty small set, but when I looked just today at extending it, none of the colors I could use some fill in were satisfied by the addition of anymore stars. I wanted to keep the weighting about the same for all colors--but lean on blue and red, and not use a lot in the middle, to get the best slopes. This gave me .253, .537, .791, 1.205, 1.521, & 2.064 for the first 6 and added .388, .439, .567, .691, 1.292, 1.733 and 1.406 for the next 7. Mix your own and add more if you choose. This is where I am at at the moment and I can transform SA110-530 to within -3, 0, +6 & 0 millimags compared to Landolt for 2 mights and about 30 frames in each filter. If I could get 100 or 300 stars, I would try that. Just not a manual operation for me. The 13 stars were an improvement over the 6, but not by a lot. I think this is because I chose the 6 with a wide color range. Don't try 6 star with a limited color range. The resuts will not be as good. (been there--tried B-V of 0.2 to 0.7 with 6 stars) It was ok, but not close enough. SA110-503 was off by 10- 85 millimags by comparison.
Hope this helps. BTW: Get a copy of George Silvas' TA (Transform Applier)--search the AAVSO web site and it should come up. Its in testing, but I have been using it a lot--version 2.12--and its really slick. No known bugs at this time, some enhancements are in the works. It will become mainstream on the AAVSO web site soon. It was put together by a team of George, Arne, Matt, Richard Sabo, Helmar, Barbara Harris, Rebecca, Mike Simonsen, Will and myself. Once you have your coeficients, you can use this to apply and generate a report to suck up by WebObs, all transformed. BTW; Gordon Sarty is working on a tool with some of the folks above to determine you coeficcients. It will be a slick tool also.
Clear Skies
Gary
It is frustrating when you have got so many stars with duplicates label. But the more frustrating thing is that in the uploaded image of NGC 7790 when you click “Load AAVSO Standard Stars” the chart generates with different star labels on every click. (see the attached images).
So as Gary said in previous post it is better to use http://binaries.boulder.swri.edu/fields/ngc7790.html
Regards,
Velimir
Hello Velimir
That's pretty bizzar. Can you add some details for HQ. This is from VSP? No it must be VPHOT, as it adds the dashes to dupes. Forgive me, I use VPHOT seldomly--Sorry Brad, old dog doing old tricks. It looks like there are two issues. In each of these plots, it designates the dupes with a different suffix. It also leaves some stars out in crowded micro fields--ie like the 147_2, 145_2, 146, and 150. Any more info you can provide will help HQ sort this out. They may already know about it. But it sure makes doing Tx more difficult. We need to continue to remove the barriers to Transformation. I know its on Arne's punch list before retirement.
Thanks for posting this.
Gary
Following Arne's advice to definitely make use of the newer photometry results in VSP
I deleted my original post (and the attachment).
Cheers,
Helmar (AHM)
Please, if at all possible, don't use the photometry table on the binaries.boulder.swri.org site. The data in VSP is far more reliable. If you want to use the same stars, IDing them from that chart, but using the photometry from VSP, that is ok.
Velimer's issue with VPHOT is an interesting one, but there must be something different between the two jpegs he gave - use of a different image, or different scale, or something - to cause VPHOT to change the stars that it marks. However, all of the marked stars are part of the standards set, so it isn't like non-standard stars are now getting labels.
Note that M67 will give you the same issue with multiple comps having the same magnitude label, and in fact, many CCD sequences will also yield duplicate labels with the chart 0.1mag granularity. The correct way to identify the stars is with their AUID, since that is unique. The various pieces of software that need to be used will eventually handle things properly.
Arne
It is a
Hi Gary and Arne,
It is a riddle for me as I do not do anything special. I have uploaded my NGC 7790 images in VPHOT and starting to create usable sequence. Very soon I realized that on the same image and the same zoom if I use the function “Clear All” and than reload “Load AAVSO Standard Stars” the labels have got new suffixes. That’s all. I have not explanation to that.
See the new attached screenshots (you can blink them)
Best regards,
Velimir
Hello Arne
Thanks for pointing this out. :-J
I compared the values in the two sources. The Std Dev of the deltas is .011 Mags and the difference in color Stdev is .006 mags. Do you have any examples of the impact of this change? Would like to know what to expect.
Gary
Hi Gary,
Was your comparison for a limited subset of the 31 stars shown on the binaries.boulder.swri.edu site, or for the whole set?
Usually, improved photometry is your best choice when calculating coefficients. With only 4 nights of observation, as per that table, you risk having systematic effects in the photometry - red stars are perhaps redder on average than for a true Johnson/Cousins system. In addition, more nights improves the photometric uncertainty on the fainter stars. Since in any list there are more faint stars than bright ones, improving their uncertainty improves the overall fit. How this corresponds to changes in your personal calibration is something that I can't predict. All I can say is that the photometry in VSP is statistically more reliable, and I strongly recommend you use the latest calibrations.
Arne
Hello Arne
I did both. I compared the new PT to a 13 star subset of the SWRI, and I compared it to all 31 of them. Of the first 13, only 9 had new data. Of the remaining 18, 16 had new data. Statistics about the same on both.
When the turmoil settles down, I may make another pass, using the new values. May keep the old vaues for those 6 stars if they cannot be duplicated for color elsewhere in the cluster. I know that the B-V= .253 does not have new values. Thats the only star bluer that about .388--kind of nice for color extension. I may run into close companion issues with some. The original 31 were pretty good at avoiding the close companion issue, except in a couple of case. I know that the first 13 did cover the .253 to 2.064 color range quite nicely. The next 8 stars did not fill in many holes. Just would add more data. Could help.
Clear Skies
Gary
I know the spreadsheet I posted was based on the old Photometry, but I did it before the new photometry was relaeased, and I have to do it all over again anyway because I added a corrector to my telescope. So now I have an opportunity to do it over with the new values and the new tools. I amy still do it using my old manual method just to see whether the results agree.
Brad Walter
Hello Brad
I agree, use your known tools. That gives an apples to apples comparison.
You would also be a great resource to test out the TDGP--Gordon's program to determine the coeficients. I know he is adding a few things. It should take a Maxim file of PT and generate coefficients.
Gary
I had to redo all the comp stars from Arne's Standard Magnitude chart (coordinates: 23:58:23.2 +61:12:25). The reason for this was that the selection of NGC 7790 stars used by PTGP 4.0 is less than 50. I started resorting these with the more limited NGC 7790 list used by PTGP 4.0, indexed by B-V, eliminated all star redder than B-V > 1.36. I also eliminated all stars in the Standard Magnitude chart that were dimmer than V-mag of 14.6. Between these two lists I was able to get 16 comp stars. I ran these through PhotomCap_14 to get me a .STAR file for AIP4WIN. I processed my 30-Aug-14 data and produced an instrumental magnitude report. I will also process last night's data on NGC 7790 to compare the Transform Coefficient btw the two nights.
James
James,
I am curious about your choice of 1.36 as the B-V limit. Are you trying to exclude stars redder than K0? NGC 7790 is a young cluster with reddening E(B-V) = 0.51. So subtracting that from 1.36 gives absolute B-V of 0.85 which is very early K spectral type on the main sequence. Arne's book lists main sequence absolute B-V for K0 at 0.82 and K5 at 1.15. I know that you normally want to avoid stars that are intrinsically very red because they tend to flicker. I don't know the generally established absolute main sequence B-V to use as a cutoff but I would think anything with absolute B-V = 1.0 would be safe. That would equate to apparent B-V = 1.51.
Arne, can you give some guidance regarding an appropriate apparent B-V magnitude cutoff to use for this cluster? We don't want to use stars that may flicker but we also want as wide a color range as possible. Also with all the photometry that has been done on NGC 7790, if a star has not been observed to vary, why not use it regardless of B-V? If the flicker is at such a low level that it hasn't been observed it probably isn't going to throw off our coefficients. That being said, I still wouldn't use a star with B-V > 2.0. because you are then in M class which are very often varying to some extent.
The other issue that I cannot address with certainty is whether some of the stars included in the sequence are actually foreground stars with much less reddening. Plotting U-B vs B-V for all 46 stars in the list of stars V=13.9 magnitude and brighter gives a curve that is consistent with all stars being in the cluster. The minimum of the A0 - F5 dip is shifted by about 0.5. from the expected non reddened B-V value of about 0.0, which is consistent with the reddening of the cluster. See attached plot.
Brad Walter.
Sorry about that. Subject says it all.
Gary
This is a pretty crowded field for some of the std stars. One might want to make measurements with a range of aperture values to see what gives the best result. I am willing to wager that the normal aperture RADIUS = 2x FWHM rule of thumb won't work, and that you end up a lot closer to aperture RADIUS much closer to 1x FWHM. In my previous photometry of NGC 7790 I ended up with aperture radii in the range of 1.0 to 1.1 FWHM on all nights I tried about 8 radii from about 0.7 to 2.5 FWHM. I also tried bunch of annulus sizes and the one that seemed to give best results for me was inner = 30 pixels and outer = 40 pixels with plate scale of about 0.63 arcsec per pixel.
When I did my last run on NGC 7790 I didn't have the corrector in the telescope so my FWHM with a big chip camera varied considerably across the field, which complicated the selection and made me do a lot of testing. I am glad that issue is now resolved.
Brad Walter, WBY
Hello All and Brad
I made a dry run on this field with Maxim/dl, to see how the apertures would fit. I got 41 of the 46. The stars that I could not get were 18, 29, 33, 44, and 46. The reason is that they were close neighbors with a star of lower number and Maxim does not let the rings overlap. I had a FWHM = 1.9 pixels and used a R aperture = 5, gap = 2 and sky anulus = 3. So I maintained the radius = 2-3 FWHM. I would rather drop stars than compromise the PT with an overly small aperture.
This is quite workable and a big help. I will try my transformations with the 41 and post how it does.
Gary
Hello
I ran the test case for NGC7790 against the 41 stars in Brad's list. I am using 3000 mm focal length and 1.6 arc sec's per pixel. I was very dubious about going to 1 FWHM, an I think a better compromise was keep the FWHM of the aperture at 2-3 pixels, and carefully pick from the lst of 41 stars, based on their color, and drop about 20% of them that are close pairs. I found certain star pairs allowed use of one, but not both stars.
I don't remember what I used for annulus settings, but was able to avoid most all stars in the annulus. Maxim does have an algoithm that will reject some level of a star in the annulus, so you don't have to be perfect.
You still get lots of great stars, good color range, and do not violate the radius of the apreture rule. I was using Maxim for PT.
Gary
Is the aperture radius rule to which you refer RADIUS = 2xFWHM?
I am going to update my list of stars and "pick list" chart. I have added the NOFS stars from the VSP comp chart that correspond to stars from the old standard list that aren't on the VSP standard chart (because stars can't be both comps and standard stars at present). I checked them against the 1999 standard values and they are consistent given the published uncertainties. The expanded pick chart will have all of the stars corresponding to 1999 standard chart except the 110 and 145 stars. The 110 star isn't included on either VSP chart and my cutoff was 13.999 V mag. The 145 star from the 1999 sequence is on the standard chart as AUID 000-BLK-022. The location is different by 0.98 arcsec but there are no close stars and I see no indication in the DSS image or in Vizier listings that it is a binary.
Brad Walter, WBY
Hello Brad
Yes, the rule I am referring to is aperture radius = 2-3 FWHM. This is particularly important for time series, if the seeing is varying. Not as important for a single image.
Gary
I hope that I am not stepping on anyones toes, but I find that when I am a lot of stars on an image to do photometry I need a picture to make sure I am selecting the right ones and I need something that uniquely differentiates between the stars. So I created a new chart that simply has the stars through 13.9 magnitude (46 stars) numbered sequentially starting with the first star on the standard sequence and going down the list through the last 139 star. For want of a better term I called these "click numbers." I made a companion Excel spreadsheet correlating click #s with labels and AUID numbers and having the magnitudes and errors listed individually in cells so that the spreadsheet can be used as a template for calculations. I know, I know, we don't have to do that anymore with our new tools, but what the heck. It might be nice to have in case you need to check results of the tools. I didn't make columns for all of the other color indexes and their error values, but that is easy for anyone to do if they want to do it. the color distribution of the stars included is
(B-V) <= 0
0
0 < (B-V) <= 0.3
0
0.3 < (B-V) <=0.5
19
0.5 < (B-V) <= 0.7
9
0.7 < (B-V) <= 1.0
2
1.0 < (B-V) <= 1.5
10
1.5 < (B-V) <= 2.0
5
2.0 < (B-V)
1
If anyone interested would check my callout placement to be sure I have assigned click numbers to the correct stars please feel free and let me know if there are any errors I need to correct. There is a triangular shaped artifact with the right angle located at the position of the 129 star,000-BLJ-973. I was in the Aladin DSS download and I didn't spend a lot of time searching around for another image that didn't have it.
See Attached. Gary, I saved the sequence table as an xls file just for you.
Brad Walter, WBY
Hello Brad
Thanks for the good work. I assume that these 46 stars are the ones down to mag 13.9 from VSP--asking for a standards chart and photometry.
Thanks for the xls charrs.
BTW: I can view all the click numbers, but when I print it out, I only get 02, 05, 09, 17, 23, and 42. The others are not designated. When I go from viewing the chart to print preview, it looses the others. Any Ideas why?
Gary